P2X4 purinoceptor signaling in chronic pain

Abstract

ATP, acting via P2 purinergic receptors, is a known mediator of inflammatory and neuropathic pain. There is increasing evidence that the ATP-gated P2X4 receptor (P2X4R) subtype is a locus through which activity of spinal microglia and peripheral macrophages instigate pain hypersensitivity caused by inflammation or by injury to a peripheral nerve. The present article highlights the recent advances in our understanding of microglia–neuron interactions in neuropathic pain by focusing on the signaling and regulation of the P2X4R. We will also develop a framework for understanding converging lines of evidence for involvement of P2X4Rs expressed on macrophages in peripheral inflammatory pain.

Introduction

Chronic pain, unlike acute pain, has no known defensive or beneficial purpose and is a reflection of aberrant functioning of pathologically altered peripheral and/or central nociceptive neural networks [46, 62, 84]. Chronic pain can be broadly categorized as chronic inflammatory pain and neuropathic pain. Chronic inflammatory pain arises as a consequence of inflammatory responses mounted by the immune system following tissue damage and generally abates after such damage is repaired [48, 58]. Neuropathic pain, on the other hand, arises because of injury to a nerve caused by trauma, infection, or pathology, and is characterized by pain that persists long after the initiating event has healed [62, 89]. In both chronic inflammatory and neuropathic pain, pain may arise spontaneously in the absence of an overt stimulus, or it can be evoked, such as in the case of allodynia (pain resulting from an innocuous stimulus) and hyperalgesia (an exaggerated pain response to a noxious stimulus). The symptoms of neuropathic pain are difficult to treat and often resistant to the current available treatments, including the potent analgesic effects of opioid drugs, which is in stark contrast to acute and chronic inflammatory pain for which there are many effective therapies. Neuropathic pain is therefore considered among the most debilitating and difficult to manage chronic pain conditions.

Previously, the predominant view was that neuropathic pain following peripheral nerve injury is the direct result of alterations in neurons and neuronal function in the peripheral and central nervous systems. Indeed, injury to a peripheral nerve instigates a series of cellular and molecular changes that directly affect neuronal plasticity, leading ultimately to altered synaptic connectivity and the reorganization of peripheral and central nociceptive circuitry [62, 83, 84, 89]. Such changes produce a dichotomous affect on pain transmission: they suppress mechanisms that inhibit pain and enhance mechanisms that facilitate pain. The net effect is a pathological amplification of sensory input and output from the spinal cord that manifests as pain hypersensitivity. The basis of these neuronal adaptations is critical to understanding the pathophysiology of neuropathic pain. However, pharmacological therapies that have been directed against cellular targets in neurons have not produced optimal therapeutic effects in patients suffering from chronic pain, in particular, neuropathic pain.

Although it is not disputed that neurons are fundamentally involved in neuropathic pain, it is now apparent that a neuron-centric approach to understanding pain is an oversimplification and does not account for the diverse network of cell types within the central nervous system. This realization is in response to a rapidly growing body of evidence that glia–neuron interactions are critical in establishing and maintaining neuropathic pain [1, 31, 36, 50]. Microglia, in particular, have emerged as key players in the etiology of neuropathic pain [36, 71, 75, 81, 82]. Microglia–neuron communication is bidirectional and considerable evidence implicates purinergic signaling as a critical molecular component [17, 23, 39, 47]. In particular, evidence is accruing that adenosine triphosphate (ATP), acting via P2 purinergic receptors, is involved in pain hypersensitivity associated with neuropathic pain and peripheral inflammation [12, 23, 24, 39]. ATP released from damaged and/or inflamed tissue directly modulates microglial functioning, and microglia in turn release a myriad of cytokines, chemokines, and neurotrophic factors that have a profound effect on neuronal function [15, 21, 36, 71, 82, 86]. In vivo imaging studies show that upon traumatic injury microglia rapidly extend processes by an ATP-dependent mechanism and converge on the site of injury [19, 32, 45, 53]. Microglia express a range of P2 purinergic receptors: of the metabotropic P2YRs, microglia express P2Y1, 2, 4, 6, and 12Rs [7, 25, 36, 61], with P2Y12R signaling mediating chemotaxis and tactile allodynia associated with nerve injury [44, 69]. By contrast, microglial expression of the ionotropic P2XR is restricted to the P2X4 and P2X7R subtypes [16, 28, 36, 75].

Activation of P2X7Rs is implicated in the microglia response to inflammation [16, 37], microglial proliferation [4, 51] and release of proinflammatory cytokines [10, 13, 15, 26, 27]. A role for P2X7Rs in neuropathic pain has also been suggested on the basis that P2X7R mRNA and protein expression are upregulated in spinal microglia follow peripheral nerve injury [43], and pharmacological blockade or genetic deletion of this receptor subtype ameliorates the development of pain hypersensitivity [9, 14, 22, 33, 34, 49, 54]. Recently, genetic variation in P2X7R function has been linked to variations in pain in mice and humans [65]. Despite accruing evidence that P2X7Rs are critically involved in both inflammatory and neuropathic pain, conclusions about the specific role microglial P2X7Rs play in these chronic pain conditions is confounded by the fact that P2X7Rs, in addition to being expressed on microglia, are localized on astrocytes and neurons in the spinal cord. Thus, involvement of P2X7Rs in chronic pain may not be specific to microglia and could instead involve activation of P2X7Rs expressed on other cell types in the spinal cord. The P2X4R, on the other hand, is emerging as a core signaling pathway in microglia that underlies the sequelae of peripheral nerve injury.

The first demonstration that P2X4Rs are involved in neuropathic pain was reported by Tsuda et al. [75]. It was found that intrathecal injection of TNP-ATP, an antagonist of P2X1-4Rs, produced a rapid and transient reversal of mechanical allodynia in rats that sustained a peripheral nerve injury 1 to 2 weeks prior. By contrast, treatment with PPADS, an antagonist of P2X1-3,5,7Rs but not the P2X4R, had no effect on mechanical allodynia when injected at the same time point after nerve injury. Based on the pharmacological profiles of these two antagonists, it was deduced that mechanical allodynia depends upon tonic activity of spinal P2X4Rs. Moreover, the time course of development of mechanical hypersensitivity was correlated with a progressive increase in spinal P2X4R protein expression and this occurred specifically in microglia. This observation was recently confirmed in the CX3CR1^+/GFP mice, in which induction of spinal P2X4R expression resulting from peripheral nerve lesion was restricted to activated eGFP expressing microglia [79]. Directly targeting P2X4Rs with intrathecal administration of a P2X4R antisense oligodeoxynucleotide prevented the increase in P2X4R protein expression and suppressed the development of mechanical allodynia [75]. An important line of evidence for the necessity of P2X4Rs comes from findings that mice lacking the P2X4R do not develop mechanical allodynia after peripheral nerve injury [72, 79]. Although the behavioral pain responses in these mice were absent, the proliferative response of microglia to peripheral nerve injury was undiminished [79], suggesting that while tonic P2X4R activation is required for maintaining peripheral nerve injury-induced allodynia, the proliferation and upregulation of microglial P2X4Rs in the spinal dorsal horn are independent. These experiments demonstrated the necessity of P2X4Rs as an active component in neuropathic pain, but do not preclude the possibility of intermediary factors being required (i.e., sufficiency had not been demonstrated). Sufficiency of P2X4R stimulation in microglia for development of allodynia is supported by evidence that direct injection of P2X4R-stimulated cultured microglia into the spinal cords of otherwise naïve animals elicits robust mechanical allodynia that is blocked by TNP-ATP [17, 75]. Collectively, the biochemical and behavioral findings indicate that activity of P2X4Rs expressed on microglia residing in the spinal dorsal horn is both necessary and sufficient to induce tactile allodynia after peripheral nerve injury.

The above results strongly support upregulation of P2X4Rs in microglia as being a critical process in neuropathic pain. Significant advances have been made towards understanding how injury to a nerve signals to the dorsal horn causing increased P2X4R expression in microglia. One such neuron–microglia signaling molecule is CCL21, a microglia-activating chemokine released from injured neurons [5, 20]. A recent study by Biber et al. [5] reported that following injury to a peripheral nerve, CCL21 expression was significantly increased in small diameter neurons in the dorsal root ganglia and in primary afferent nerve terminals in the spinal dorsal horn, but not in microglia or in astrocytes. Early intervention with intrathecal delivery of CCL21 neutralizing antibody before and for the first 3 days after nerve injury significantly attenuated tactile allodynia throughout the 14-day experiment in wild-type mice; this early preemptive effect argues for involvement of CCL21 in the initiation of neuropathic pain. In accord, mice that lack neuronal CCL21 did not develop tactile allodynia following nerve injury and failed to upregulate microglial P2X4R expression. Stimulation of microglia in culture with recombinant CCL21 induced a rapid and concentration dependent increase in P2X4R expression, suggesting a direct effect of CCL21 on the microglia. This notion is corroborated by observations that intrathecal injection of CCL21 to nerve injured CCL21 deficient mice induced long lasting allodynia and upregulated P2X4R expression. The pronociceptive actions of CCL21 are strictly dependent on P2X4R function because injection of this chemokine in P2X4R knockout mice had no effect on tactile allodynia. Taken together, these findings reveal CCL21 as a key neuronal molecule that functions as an upstream activator of P2X4R expression in spinal microglia that is required for the induction of neuropathic pain.

Several other signaling elements have been implicated in the upregulation of P2X4Rs following peripheral nerve injury. For example, the cytokine interferon γ has been identified as a signaling molecule that drives P2X4R expression and through which resting spinal microglia transform into an activated state [73]. Release of tryptase from activated mast cells has also been found to increase P2X4R in microglia [85]. Likewise, microglia grown in culture in the presence of the extracellular matrix molecule fibronectin show a marked increase in P2X4R levels and a consequent enhanced Ca²⁺ response to ATP stimulation [52]. Fibronectin expression was elevated in the spinal cord of nerve injured animals and blockade of fibronectin receptors suppressed both the increase in P2X4Rs and tactile allodynia [52, 76–78]. Fibronectin induced increase in microglial P2X4R requires activity of the tyrosine kinase Lyn and downstream activation of intracellular signaling pathways involving phosphatidylinositol 3-kinase (PI3K)-Akt and mitogen-activated protein kinase kinase (MAPK kinase, MEK)-extracellular signal-regulated kinase (ERK), which have distinct roles in the up-regulation of P2X4R expression in microglia at the transcriptional and post-transcriptional levels, respectively [77]. Thus, several critical signaling elements necessary for the upregulation of P2X4R in microglia following nerve injury have been identified, but how these elements are causally connected and whether they encompass the entirety of the necessary pathways remains an open question.

P2X4Rs are known to undergo rapid constitutive internalization and reinsertion to the plasma membrane [6, 30, 59, 67]. In microglia, internalized P2X4Rs are targeted to lysosomes and constitutive retrieval of these receptors from the plasma membrane regulates the proportion of P2X4Rs on the cell surface [56]. Lysosome exocytosis, stimulated by incubating microglia with the Ca²⁺ ionophore ionomycin, leads to an accumulation of P2X4Rs on the cell surface and enhances P2X4R-mediated currents [56]. Cell surface targeting of P2X4Rs by lysosomes has recently been shown to involve the chemokine CC2 [68]. It is well established that injury to a cell induces lysosome exoctyosis [35, 38, 57, 66]; thus, it is conceivable that cell damage resulting from injury to a nerve might mobilize P2X4Rs from lysosomes to the microglial cell surface.

P2X4Rs expressed on the cell surface are capable of being stimulated by ATP. Activation of these receptors instigates a series of conformational changes that allow cations, such as Ca²⁺ and Na⁺, entry into the cell through a non-selective channel [11]. The recent isolation of the P2X4R crystal structure has illuminated many atomic details about its extracellular domain structure, the putative ATP binding site, transmembrane regions, and ion permeation pathway [41]. The C-terminus of the P2X4R in particular appears important in agonist-induced desensitization [29], binding of phosphoinositide PIP₂ [3], and constitutive internalization [30, 56, 59]. Time-lapse imaging of ATP induced structural changes in the P2X4R has revealed two distinct activated conformations: in the presence of extracellular Ca²⁺, ATP stimulation caused the P2X4R to open a non-selective cation-permeable channel, but in the absence of extracellular Ca²⁺, the receptor formed a transmembrane pore (i.e., pore dilation) that allowed passage of large molecules [64]. P2X4R expressed on microglia and macrophages have recently been reported to possess this ability to function as both an ion channel and a pore [2, 63]. Although the biochemical and biophysical ion channel properties of the P2X4R are well characterized and widely investigated in microglia, the significance of this dual function, in particular the role of P2X4R pore dilation in the context of neuropathic pain remains to be elucidated.

The first clues that activity of P2X4Rs expressed on microglia alters fundamental spinal nociceptive processing stem from a series of experiments by Coull et al. [17], who demonstrated that intrathecal administration of ATP stimulated microglia produced a depolarizing shift in E_anion and converting GABA-evoked responses from hyperpolarizing to depolarizing in spinal lamina I neurons. The altered inhibitory responses initiated by P2X4R stimulation in microglia unmasks low threshold input to lamina I, thereby driving the phenotypic switch in their response properties. This phenotypic switch caused spinal lamina I neurons to relay innocuous mechanical input, increase discharge to a noxious stimulus, and display spontaneous activity [18, 42]. The changes in output of the lamina I neurons may be interpreted as providing a neural substrate for neuropathic pain. Because microglia are not directly involved in the spinal neural network that process and relay nociceptive information, Coull et al. [17] reasoned that microglia must signal to lamina I neurons to affect neuronal hyperexcitability. Indeed, several key findings pointed to brain-derived neurotrophic factor (BDNF) as being the critical microglia–neuron signaling molecule: (1) function blocking antibody against trkB, the cognate receptor for BDNF, and the BDNF-sequestering TrkB-Fc fusion protein prevented mechanical allodynia evoked by administering P2X4R-stimulated microglia; (2) treatment of cultured microglia with siRNA against BDNF prevented the effects of intrathecally administered microglia on lamina I neurons and on the microglia-elicited pain behaviors; and (3) P2X4R stimulation by ATP caused release of BDNF from microglia in culture, a response prevented by TNP-ATP or by knocking down expression of BDNF with siRNA [17]. Taken together, the most parsimonious explanation for these findings is that P2X4R stimulated microglia signal to spinal lamina I dorsal horn neurons causing aberrant nociceptive output and that the critical microglia–neuron signaling molecule is BDNF. The conclusion that P2X4Rs control BDNF release from microglia is congruent with observations in P2X4R-deficient mice that BDNF accumulates in spinal dorsal horn microglia following nerve injury and that microglia isolated from these mice show impaired ATP-evoked BDNF release [79]. However, it has yet to be shown that resident microglia in the spinal cord respond to nerve injury by releasing BDNF. In mice, where BDNF was genetically eliminated from small diameter sensory neurons, inflammatory pain was significantly attenuated but development of allodynia following nerve injury was unperturbed [87]. This finding is consistent with primary afferent derived BDNF not having a causal role in the sequelae of neurpathic pain arising from injury to a peripheral nerve.

We recently addressed the gap in understanding how activation of P2X4R leads to the release of BDNF from microglia [70]. We demonstrated in cultured microglia that influx of Ca²⁺ through the P2X4R is a critical intracellular step linking stimulation of these receptors to the activation of p38 MAPK. We discovered that p38 MAPK is required for the SNARE (soluble N-ethylmaleimide-sensitive factor attachment protein receptor)-dependent vesicular release of pre-made BDNF and also increased transcription and translation of BDNF. By demonstrating that p38-MAPK is a cellular intermediary in the release and synthesis of BDNF evoked by stimulating microglial P2X4R, we provided a unifying mechanism for observations that ongoing expression of neuropathic pain behaviors requires both the activation of P2X4R [17, 75, 79] and p38-MAPK [40, 74, 88]. These findings place p38 MAPK in the core P2X4R-BDNF signaling cascade and implicate p38 MAPK activation and BDNF release as an essential step in microglia–neuron communication leading to nerve injury induced pain hypersensitivity. Likewise, activation of P2X7Rs or P2Y12Rs leading to the development of pain hypersensitivity engages distinct intracellular signaling pathways in microglia that converge onto p38 MAPK [15, 44]. Thus, p38 MAPK is a cellular intermediary and its activation is a critical point of convergence for P2X4R, P2X7R, and P2Y12R signaling in neuropathic pain. The significance of this convergence in microglia signaling and the upstream, as well as, downstream components of these signaling pathways in aberrant spinal nociceptive processing following nerve injury remains to be fully elucidated.

A similar link between P2X4R and p38 MAPK signaling has recently been uncovered in peripheral macrophages [80], which like microglia are immune cells that constitutively express functional P2X4Rs [8, 80] targeted to the cell surface by lysosome trafficking [56]. Rassendren and colleagues demonstrated that stimulating P2X4R expressed in macrophages triggers Ca²⁺ influx and p38 MAPK phosphorylation that drives the production and release of prostaglandin E2 [80], a principle substrate for inflammation that elicits pain hypersensitivity through sensitization of peripheral nociceptors [55, 60]. Genetic deletion of the P2X4R abrogates P2X4R-mediated release of prostaglandin E2 from macrophages [80] and prevents pain hypersensitivity in response to inflammatory challenges [72]. Thus, these discoveries highlight a novel role for P2X4R in inflammation and represent a significant conceptual advance that extends involvement of P2X4R in chronic pain beyond neuropathic pain to encompass the pathoetiology of inflammatory pain.

In conclusion, the P2X4R has emerged as a core microglia–neuron signaling pathway that is necessary for ongoing tonic expression of tactile allodynia following nerve injury. Converging lines of evidence also point to activity of P2X4Rs as being a locus through which spinal microglia and peripheral macrophages contribute to neuropathic pain and inflammatory pain, respectively. In both cell types, the cellular and physiological consequences of P2X4R activation are mediated by influx of Ca²⁺ through the P2X4R and activation of p38 MAPK (Fig. 1). This mechanistically identical signaling step results in the release of distinct signaling molecules from microglia (BDNF) and peripheral macrophages (prostaglandin E2) that alter neuronal nociceptive output and translates into pain hypersensitivity. Thus, identification of the upstream and downstream components of the P2X4R signaling pathway has provided important new information about the fundamental central and peripheral mechanisms that underlie pain hypersensitivity subsequent to nerve injury and inflammation. The P2X4R is therefore a promising therapeutic target for the treatment of neuropathic pain and inflammatory pain.

Fig. 1.

P2X4R-p38 MAPK is a core signaling pathway through which spinal microglia and peripheral macrophages contribute to neuropathic pain and inflammatory pain. a Following injury to a peripheral nerve, expression of P2X4R is upregulated in microglia that reside in the spinal dorsal horn. Several factors have recently been implicated in the upregulation of microglial P2X4R: the fibronectin-Lyn kinase signaling cascade, cytokine IFN-γ, tryptase released from mast cells, and the chemokine CCL21. Influx of Ca²⁺ through ATP stimulated P2X4R activates p38-MAPK and drives the synthesis and SNARE-dependent release of BDNF. Acting on its cognate receptor, trkB, BDNF released from microglia signals to dorsal horn lamina I neurons to downregulate the K⁺-Cl⁻ cotransporter KCC2, resulting in aberrant nociceptive output that underlies pain hypersensitivity characterized by hyperalgesia, allodynia, and spontaneous pain. b Like microglia, peripheral macrophages constitutively express P2X4R. Inflammatory challenges that activate P2X4R on macrophages trigger Ca²⁺ influx and p38 MAPK phosphorylation. In macrophages, the P2X4R-p38 MAPK pathway is a signaling hub that activates cytosolic PLA2 which liberates arachidonic acid (AA) resulting in a cyclooxygenase (COX) dependent synthesis and release of prostaglandin E2 (PGE2), a key substrate that sensitizes primary sensory neurons and leads to hyperexcitability of peripheral nociceptive pathways that is a hallmark of inflammatory pain